What is Text-to-speech?
Text-to-speech (TTS) turns written text into spoken audio, and is the voice a voice agent uses to speak its replies.
In more detail
Modern neural TTS sounds natural and can stream audio while the text is still being generated, which cuts delay on calls. The factors to compare are voice quality, time to first audio, language and accent coverage, pronunciation control for names and numbers, price per character and licensing terms. Long numbers, dates and addresses often need special formatting to sound right, and a voice that reads an order number fine in a demo can mangle it on a noisy phone line. If you clone a voice, get written consent from the speaker. Some jurisdictions regulate synthetic voices on calls, particularly for outbound marketing, so check local rules before you launch.
Example
A parcel carrier's voice agent must read tracking numbers back to callers. The team formats them in groups of three characters with short pauses, so customers can write them down.